Dogs and humans have worked in partnership throughout history thanks to dogs' unique capability of detecting signals in human voices or gestures and learning from human inputs. Traditional canine training methods rely solely on subjective visual observations made by trainers. We propose a canine body-area-network (cBAN) to incorporate context-aware sensing with objective detection algorithms to augment the sensitivity and specificity of human trainer's awareness of the dogs they are training. As an initial effort, we developed a Bluetooth-based wireless infrastructure and tested inertial measurement units as cBAN sensor nodes to electronically assess the posture of the dogs. As a result, we were able to optimize the sensor locations and distinguish different postures using the distinct patterns in the measured angles.

Canines and humans have lived together for many thousands of years, to our mutual benefit. In addition to providing companionship, dogs can perform critical roles, such as assisting humans with medical concerns, searching for lost individuals, and detecting substances by scent. Researching how technology might be designed for canines has the potential to significantly improve the lives of both dogs and humans. We draw upon the extensive foundations and literature in human-centered computing to identify and adapt models and methods that are relevant for canines. Our work surveys the landscape of canine-centered computing and canine psychophysics, and generates a framework and set of guidelines to help inform the requirements, design, and evaluation of systems for canines. Our principal aim is to invite and challenge human computer interaction (HCI) researchers to contribute to the field of canine-centered computing, and we conclude with a call to action in this promising nascent field.

This paper introduces two novel algorithms for learning behaviors from human-provided rewards. The primary novelty of these algorithms is that instead of treating the feedback as a numeric reward signal, they interpret feedback as a form of discrete communication that depends on both the behavior the trainer is trying to teach and the teaching strategy used by the trainer. For example, some human trainers use a lack of feedback to indicate whether actions are correct or incorrect, and interpreting this lack of feedback accurately can significantly improve learning speed. Results from user studies show that humans use a variety of training strategies in practice and both algorithms can learn a contextual bandit task faster than algorithms that treat the feedback as numeric. Simulated trainers are also employed to evaluate the algorithms in both contextual bandit and sequential decision-making tasks with similar results.